stereoEEG

Repository to house code for a collaborative project analyzing stereoEEG data during anesthesia induction, anesthesia emergence, and sleep for NINDS-R01NS113366: The role of dynamical criticality in human perception

Downloading data from iEEG.org

Process and save .mat file for each session with:
process_ieeg_data;
If tone response data is available for the session, make sure the 'toneResponse_Induction' subfolder is in the specified 'data_dir/subject' path so that it is included in the .mat file.

Add brain segmentation labels to each sEEG contact

Run addSEEGlabels; (or addSEEGlabels_rave; for newer subjects) to add labels to the channel_labels variable in the .mat file. Uses labels found in ITKSnap CSV file (addSEEGlabels) or exported from RAVE (addSEEGlabels_rave). This is currently only relevant for the induction sessions.

Sync behavioral-sEEG data

If tone response data is available for the session, run toneResponseSync;. This adds the tone on, tone off, button press, button release timestamped events to the session annotations.

Analyze behavioral data

Run toneResponseAnaly; (requires functions in Smith2004 subfolder) to estimate probability of response based on response data following analysis of Smith et al 2004 J Neurosci. Estimate of time of loss of consciousness (LOC) follows similar criteria as Purdon et al 2013 PNAS.

Identify channels to include in analyses

Run ibad = notSEEGchannels(channel_labels,wm,bad_channels); to return indices of channels not to include in subsequent analyses. Removes non-sEEG channels (e.g. scalp EEG channels), channels labeled as white matter if set second argument to 1, and any previously identified 'bad' channels listed in cell array in third argument.

Qualitative analysis: time domain

Load .mat file for session. Then produce scrolling time plot with:
timePlot(data,sample_rate,[],[],1,annotations,channel_labels(:,1));.

Qualitative analysis: frequency domain

Load .mat file for session. Then produce power spectral density (PSD) plot with:
pwelchPlot(session.data,session.sample_rate,[0.5 500],5,1,500,session.channel_labels(:,1));
Identify bad channels by clicking on outlier PSD with Edit Plot tool and finding channel name:
get(gco,'Tag')

Stability analysis

Replicating analysis of Solovey et al 2015 J Neurosci (Figures 2, 3, 4). Load .mat file for session. Then perform analysis with:
seegStability(session);. All available sessions can be batch processed using:
seegStabilityBatch; which will generate a plot for each session and save a .mat file for population analysis.
After loading that population .mat file containing the structure stability, the population analysis can be run with:
seegStabilityPop(stability);

Time-frequency analysis

Computes both average power and average criticality in each time x frequency bin. Load .mat file for session. Then perform analysis with:
seegTF(session);.

Criticality-frequency analysis

Computes distribution of mode criticality vs mode frequency for a specified pair of time windows (awake vs anesthetized) in a session. Load .mat file for session. Then peform analyis with: seegCF(session,tawake,tanesth);.